I want to push back on this post. I think sadly this post suffers from the same problem that 99% of all legal advice that people receive suffers from, which is that it is not actually a risk analysis that helps you understand the actual costs of different decisions.
The central paragraph that I think most people will react to is this section:
Being involved in litigation, even as a totally blameless witness—or even a perceived witness who in fact has no relevant knowledge at all—is expensive, time consuming, emotionally taxing, and unpleasant. Even cheap lawyers cost hundreds of dollars an hour these days and often bill in increments of .1 hours for their time. Anyone who gets caught up in court proceedings can expect to pay such a lawyer (or have their employer do so) for many hours of time to help them produce documents and communications (or formally object to having to do so) and then prepare them to be grilled for many more hours by another well-paid professional interlocutor with goals and motives at best orthogonal to their own, if not outright hostile.
Sadly, this post does not indicate what the actual expected time a witness might be expected to spend in litigation is.
I don’t currently know this number, but my guess is that less than 10% of any EA leaders who would end up writing publicly about this would end up being involved in any substantial amount of legal proceedings. My guess based on doing some independent legal research myself is that the actual cost of being involved in this kind of litigation is substantial, but not massive. Most likely a few thousand dollars in lawyer fees (for the simplicity of calculation, let’s say $10k), and probably a few days of being involved in court proceedings (let’s say for simplicity 5 days, spending like 5 hours on each of those days).
Finishing this very rough estimate, my guess is the actual expected cost of writing publicly about this is around 0.1 * $10k + 0.1 * 25 hours = $1k + 2.5 hours. My guess is this is well below what most EA leaders would be willing to pay to be able to speak openly and publicly about this, and would at most be a minor consideration on whether to actually say anything on this topic.
That said, I am not a legal professional, and my estimates might be totally wrong here. I would really appreciate a legal analysis that actually estimates the cost here, since I think as written, it seems totally plausible to me (and indeed more likely than not) that the expected cost of speaking up is pretty minor, though it is also plausible that the expected cost could be major, and settling which one it is is the primary question that legal analysis should aim to answer, in my opinion.
“Lawyer” is traditionally a high-status job, and most people look to lawyers with some degree of deference and authority. When a lawyer says “don’t do X, you might get sued”, people usually listen. But lawyers are institutionally trained and personally incentivized to be conservative, largely because of asymmetric justice—if you get sued and the lawyer didn’t speak up, they could be blamed, while lawyers are almost never blamed for shooting down an otherwise high-value idea out of inaccurate risk assessment. Just like airports and missing a flight, an organization that always listens to lawyers is being more cautious than the optimum.
I consider your attempt at a quantified expected cost analysis a helping hand, not pushback, and I appreciate it.
Accepting it as a data point, a few quick points in response:
Your comment only addresses the self-interest angle, which was a relatively small part of my post. It (understandably) ignores the impacts on others and the systemic impacts that I tried to highlight, which I don’t think can be disentangled from the self-interest analysis so easily. I’m not sure those additional impacts are amenable to quantified expectation analysis (though I’d be happy to be proven wrong on that), but we shouldn’t just ignore them.
I think your numbers are low at the outset, but I don’t think any tweaking I’d do would cause us to be off by an OOM. That said, I think you’ve established a floor, and one that only applies to a witness with no potential liability. Accepting your numbers for the sake of discussion, the time estimates sit at the bottom of towering error bars. And that’s assuming we’re talking about an individual and not an organization that might have orders of magnitude more documents to review than an individual would, attorney-client privilege and other concerns that complicate things greatly and require a great deal more attorney time, and potentially many individuals involved as witnesses, with all the associated costs for each one.
The expectation analysis also (again, understandably) only addresses time and money and does not address mental and emotional impacts. Such things are obviously difficult to quantify, but they are worth taking seriously. The legal system fails to address these issues at basically every turn, and causes a great deal of suffering as a result.
On the point about lawyers being overly conservative, I think from a 10,000-foot view assessing emergent trends across all legal advice given in all situations, I agree with this. But when you dig into individual instances of legal advice, the idiosyncrasies of each situation make applying a heuristic based on emergent macro-trends difficult to justify. I recognize that’s not a solution to what I agree with you is a problem, but I think it’s an important part of the problem to recognize in the context of this discussion.
I consider your attempt at a quantified expected cost analysis a helping hand, not pushback, and I appreciate it.
Thank you! I am glad.
Some further thoughts in response:
Your comment only addresses the self-interest angle, which was a relatively small part of my post. It (understandably) ignores the impacts on others and the systemic impacts that I tried to highlight, which I don’t think can be disentangled from the self-interest analysis so easily. I’m not sure those additional impacts are amenable to quantified expectation analysis (though I’d be happy to be proven wrong on that), but we shouldn’t just ignore them.
I do admit to failing to understand your case that the impact on others and the systemic impacts are expected to be negative, not positive.
It seems that our prior should be strongly that making the courts have more relevant information will overall cause them to make better decisions, and while you list some plausible scenarios of negative impact, the scenario of “someone says something true publicly that then gets used in the lawsuit to help identify the circumstances and details of the crimes that appear to have been committed in a way that both makes justice more likely to be achieved and saves everyone time” seems by far the more likely outcome.
My guess is that if you had a class action lawsuit by the FTX credits or people who deposited their money into FTX that they would pretty strongly prefer more people go give more information on FTX publicly, since that will likely make their job easier, not harder, and as such I think the sign of the secondary effects here points strongly in a positive, not negative, direction.
I think your numbers are low at the outset, but I don’t think any tweaking I’d do would cause us to be off by an OOM. That said, I think you’ve established a floor, and one that only applies to a witness with no potential liability. Accepting your numbers for the sake of discussion, the time estimates sit at the bottom of towering error bars. And that’s assuming we’re talking about an individual and not an organization that might have orders of magnitude more documents to review than an individual would, attorney-client privilege and other concerns that complicate things greatly and require a great deal more attorney time, and potentially many individuals involved as witnesses, with all the associated costs for each one.
I agree that if you have potential liability the costs might be higher, though it’s a bit unclear what our moral imperative in that case should be (like, depends on whether your liability here is the result of kind of crazy laws or not).
I think there are definitely some individuals who have more at stake here, and where it probably makes sense to be more quiet, though I would be quite surprised if this includes most of the EA leadership.
I feel a bit confused about how much organizations could get involved in this situation, and would appreciate more datapoints on what it means for an organization to get involved in this at a level higher than individual witnesses. I understand the obligations that individuals are under if they are called in for testimony, but don’t understand the obligations that organizations could be under in a case like this.
The expectation analysis also (again, understandably) only addresses time and money and does not address mental and emotional impacts. Such things are obviously difficult to quantify, but they are worth taking seriously. The legal system fails to address these issues at basically every turn, and causes a great deal of suffering as a result.
I agree that there are substantial costs here, but I also want to talk about the symmetric costs on the other side that strike me as substantially greater. My sense is most people will suffer a good amount more stress from trying to be quiet about this topic for many years, and will incur much greater costs from there being a lot of latent concern and anxiety for many years as people stay quiet, and that the mental and emotional impacts of that will be much greater than the impact of being involved in the proceedings itself, at least in the aggregate (I do think some individuals will have a pretty bad time with the court system, but I think approximately everyone in EA will have a pretty bad time if discussion around this topic stays guarded for many years).
Criticism of the US model of a strongly “adversarial” justice system is definitely valid; it wastes a lot of resources on inefficiently resolving certain sorts of disputes. However, the bankruptcy case is in a US federal court, and at least most of the key players are within the jurisdiction of said court. And I don’t think you usually get to a fair, equitable, or efficient outcome by not asserting your rights and interests when the other side is going out with guns blazing. The system tends to give both sides some good cards and some cards that suck; getting an outcome that is fair to all involved generally involves giving up your good cards in exchange for the other side giving up theirs too.
There is a parallel system of alternative dispute resolution (ADR) like mediation which tends to be much more efficient and more likely to produce outcomes everyone can live with. But it requires give and take by both sides, and right now the debtors are not in a position to engage with it. They have much more pressing issues to deal with in the next few months, like trying to track down $10B in missing crypto before it ends up untraceable or outside the practical control of a US court.
In my view, many organizations do have a moral obligation to return some of the funds—and an ADR process is the best way to do that. It is reasonable for the organizations to ask for dismissal of any claimed liabilities they should not morally have to bear in exchange for that.
Most litigation of note will be against organizations—like Willie Sutton didn’t actually say about banks, because that’s where the money is. To me, the likely story is that corporate counsel is advising organizations about the desirability of directing their employees to stay mum to protect the organization’s legal interests. In fact, organizational counsel does not represent the employees (which is a good thing to remember if you ever are involved with a lawyer at work!) That directive may incidentally benefit the employees and directors, though.
It is really difficult to give specific examples of the corporate/individual interplay because we don’t know what the identified legal risks are. But two general principles: first, an employer is generally liable for an employee’s actions in the scope of the employee’s employment. What “in the scope” means is pretty fact-specific. As an extreme example, we have the Department of Justice arguing at the moment that denying allegations of sexual misconduct from decades ago is in the scope of employment of the (former) U.S. president.
Second, corporations lack minds and cannot “know” anything other than through affiliated people. So for instance, if the CEO of Organization X comes on here and writes that he knew A, B, and C (at a time they were CEO), that will probably establish that Organization X knew A, B, and C. That might or might not be the case if a more junior employee said that they knew A, B, and C but higher-ups did not know. All that is to say that the speech of affiliated individuals tends to have effects on the legal interests of the organization too.
Unfortunately, because none of us know what is being discussed by the people involved, we are limited to speculation about what might be going on. Speculation is a poor substitute for information, but it can’t leak knowledge (that the speculator does not have) or be attributed to a person or organization (with whom the speculator has no affiliation). I know that is not very satisfying.
Sorry for the delay in responding here, and apologies in advance for keeping this one short.
Your position appears to assume that speaking outside the court process after the fact is likely to add accurate information to the public record and help everyone get to the truth. One of the main thrusts of my OP is that this is not the case. Instead, speaking outside the process after the fact is more likely to create confusion instead of clarity, along with all the other negative effects I walk through. I’m speaking strictly from my experience and in the abstract here, but I think that’s where we disagree. Even information provided carefully, with counsel, under the controlled circumstances of the judicial process can be problematic, but it at least has the benefit of being relatively structured and under oath. It’s much more likely to help everyone (including class action attorneys) get as close to the truth as possible.
I’m not saying there are not potential upsides of speaking out—including to the mental and emotional well-being of folks who want to speak out and have been advised to stay quiet. I’m just stating my view that those upsides are outweighed by the potential downsides, not only to the speaker, but to folks they might speak about and the process in general.
Re the question about how organizations could get involved, I don’t pretend to know the specifics here and wouldn’t comment if I thought I did, but on a general level Jason covers this in his response.
I appreciate your take here, but would also take bets at pretty large odds that if we were to ask the relevant creditors and defendants that they would prefer people to speak openly here, and that this guess of yours is inaccurate.
I think I just don’t understand the basic argument by which information here will harm people, especially if that information provides accurate evidence about actual wrongdoing that happened, or provides substantial evidence about who might be relevant witnesses to the case, by modeling the social dynamics around FTX.
The prior in favor of “accurate information helps people come to more accurate beliefs” just seems so much stronger here than the relatively vague and abstract arguments you are providing.
The prior in favor of “accurate information helps people come to more accurate beliefs” just seems so much stronger here than the relatively vague and abstract arguments you are providing
A simple model, which is roughly what I interpret legal experts to mean when they say that the US justice system is intentionally designed as adversarial, is something like: the probability that a judge finds you guilty (pj) is the product of the true probability that you are guilty (pt), the bias on evidence introduced by the prosecutor (bp), and the bias on evidence introduced by the defender (bd):[1]
pj=ptbpbd
We observe that pj=pt↔bp=b−1d i.e. the optimal truth finding strategy is when both sides are equally biased in their own favor. And importantly to your point, one side being unbiased results in the judge making a worse decision (assuming the other side is biased).
Furthermore, the equilibrium of this game is when sides are maximally biased in their own favor (since otherwise the more biased side would disproportionately win). Knowing this, the US legal system encourages both sides to be maximally biased, since this results in the most accurate judgments.
I would be interested to hear from others if this seems accurate, as it’s not something I know a lot about.
FWIW, I’d happily take the other side of a bet from you here, if one could construct a wager that would settle this. But I don’t think that’s possible, which I reckon is part of the problem.
And while I’ll confess to being a little miffed at having my own prior (updated over 15+ years directly observing analogous circumstances) downgraded to a ”guess” in your comment, while your own remains on its linguistic perch, I don’t think our continuing to go around on this will be useful. I’ll propose we “settle” it by my buying you a beer (or a cup of coffee, if you prefer) the next time we’re in the same city (and I’ll draw my home circle out to include NYC, if you happen to come through), and we can spend the time discussing the many things I’d wager we agree on.
For the non-lawyers, I’d add that two of the factors you may not be thinking about are (A) the cost of the attorney gathering enough information to make even a semi-reliable estimate of expected costs; and (B) the risk of client error from various sources.
As a general matter: Client officials are not lawyers and may not understand what information is relevant to bring to the lawyer, even with some guidance they will need to err on the side of overproduction. So the lawyer often needs to review a lot of irrelevant information to better ascertain the actual risks involved, which raises costs.
Second, client officials have their own motivations which may run at cross purposes to the lawyer making an accurate risk assessment. Let’s say—hypothetically—that an employee did something that is in retrospect really inadvisible and that exposed the organization (and possibly themselves) to significant liability. That the attorney represents the corporation and not the employee, and should be explaining that to employees if there is any risk of a conflict of interest between the employee and the corporation. How likely is it that the employee is going to ’fess up with minimal prompting?
I think you’re quite right to be asking these questions (although I think your estimate of expected cost is significantly too low for many leaders and organizations).
A good lawyer (whose client is willing and able to pay for good advice) generally does not issue commands. They take time to understand the client’s interests, objectives, and tradeoffs and then provide professional advice explaining risks/tradeoffs to guide the client as to how to move forward. The choice of objectives belongs to the client as long as those objectives are lawful. Unfortunately, the client can’t tell the rest of the community about what risks were considered without blowing the attorney-client privilege and providing very valuable information to the other side.
The original comment is probably vague on the risk/benefit tradeoff in part because Tyler is not any organization’s lawyer and does not have the information needed to assess each organization’s risk specifically. I took his main point to be that non-lawyers would be prone to underestimate the various costs of speaking out, and that it wouldn’t be appropriate to draw any inferences from anyone’s decision not to do so. I don’t think he was trying to provide legal advice on whether anyone should speak out. The benefits of speaking out are non-legal benefits that are really for the client to weigh given its objectives, and have been discussed elsewhere at some length on this forum.
At Oliver’s request I dug into the Madoff case a little, to see if there was anything we could learn from.
Epistemic status: I am not a lawyer and I spent 2.5 hours on this, one of which was unproductively wrestling with the court records database, the worst database UI in the history of databases.
The following are court cases I found related to Madoff Investments:
Bernie Madoff pled guilty and never went to trial. It wasn’t even a negotiated plea bargain, just a straight guilty please (probably so he didn’t have to testify against anyone)
His two top lieutenants took plea deals and never went to trial (but did testify elsewhere).
5 other employees had a joint criminal trial. The Madoff case broke in 2008, they were indicted in 2010 (not all together), the case went to trial in 2013, and took 8 months to complete. Appeals continued until 2016.
The main trustee in charge of clawback filed “over 1,000” civil suits aimed at clawing back money. None of the ones I found in the news went to trial, although some got very close. I probably identified <20 cases, primarily the largest ones.
The lawsuits were (reportedly) aimed at people who took more money out than they had ever put in. The ones that made the news typically had a component of “knew or should have known Madoff was a fraud” and encouraging others to invest. However the money recovered went to a general fund, not the people they specifically were responsible for bringing to Madoff.
I found a handful of other lawsuits brought against the same entities from the large set of civil suits. None of these appear to have gone to trial, although they got very little coverage so it’s hard to say.
So that’s one trial. I spent an hour trying to get a witness list and failing, and am now hoping someone reads this and Cunningham’s laws me. You’re looking for S.D.N.Y., No. 09-mag-2484 or 1:10-cr-00228, transcripts are $6+ a page to get quickly. You might be able to do better going to the courthouse in NYC.
In the meantime:
That trial had “over 40” witnesses. Let’s call it 43.
The trial lasted 8 months, although not ever day had witness testimony.
Of those witnesses I found the following:
2 Expert witnesses/investigators who clearly opted in
That leaves 30+ witnesses unaccounted for, which isn’t really enough to draw interesting conclusions, especially without knowing the denominator of who could have been called.
3 character witnesses (which is presumably voluntary)
1 contractor
Defendants: 2
Expert Witnesses: 3
Non-enron character witnesses: 11
The prosecution also shared its prospective witness list. It has 31 names, of which 23 testified, plus another 2 who weren’t on the original list. Of those two, one was an FBI agent for whom this was part of his job, and another was an Enron employee brought in as a rebuttal witness.
Enron produced a lot of trials, of which this is only one. There were so many trials Wikipedia didn’t even bother to list all of them (example), plus the senate hearings, extradition trials, appeals, etc. This is a very narrow slice of who had to testify about Enron.
But all of the witnesses either opted in (as character or expert witnesses, or investigators), worked for the company or person that committed crimes, or worked for the auditing firm that missed the crimes. None of these seem equivalent to the people I imagine you are talking to or about here. Could you give some specific examples of witnesses in positions similar to those people, in cases similar to FTX’s likely case? I’m happy with a very broad definition of similar, and I would absolutely believe they exist and I missed them. I’m a non-expert who has looked at two cases so far and both are missing crucial aspects of EA’s relationship to FTX. But I’m clearly not going to find them very fast.
Full disclosure: habryka is paying me for this research. He didn’t review either comment before publication but I did talk to him about my findings as I was working. I meant to include this in the first comment but the “(paid)” in “At Oliver’s (paid) request” got lost in an edit somewhere.
The scope of what the prosecution has to prove in a criminal prosecution can be surprisingly narrow. Specifically, the prosecution needs to prove the elements of the charged offenses—not the entire scope of the criminal enterprise. Assuming there is a conviction, other information can come in at sentencing where the proof is much simpler. In contrast, in civil litigation, you need a much more complete picture to establish who the bankruptcy estate might be able to get money from.
As this article notes, the Madoff trustee’s firm ran up $1.4B in legal fees as of last year on civil litigation. You note that they filed about 1,000 civil lawsuits—I suspect there were significantly more targets than that because the trustee viewed litigation as a last resort when settlement negotiations failed. I have no way of knowing how many different individuals and entities got hit with depositions and discovery demands in those 1,000+ cases . . . but I would suspect the average number of depositions and discovery demands in a filed civil case is greater than one each. The trustee had to be doing something substantial to justify the court approving all those fees . . .
I suspect there was less civil litigation around Enron, but for reasons that aren’t helpful here. Enron was a publicly-traded corporation, and IIRC the bulk of the losses were equity (people who held Enron stock that was overvalued due to fraudulent financial reports). Neither Enron nor those who lost money on Enron stock would be expected to have a claim against people who traded in its stock without knowledge of the fraud. Among other reasons, Enron itself wasn’t a party to those transactions, so its bankruptcy estate would have no clawback claims against third parties who sold Enron stock to other third parties for more than it was ultimately worth.
I don’t think this supports any conclusions yet, it’s two cases and I only got a full witness list for one of those. I would love to see any examples of witnesses who were in roughly the position EA leadership or members (allegedly) are- people who received money for charitable purposes, from someone who earned it through fraud and needs to pay it back, but weren’t close relatives or could otherwise be considered tightly aligned with their funder. Or someone like Holden, whose org AFAIK didn’t receive any FTX money at all.
(FTR I think Holden waiting until his takes are cold is entirely reasonable, I think the speed of response on the EAF is overheated, but given the trials could take over a decade, establishing the risks of speaking candidly before they are entirely resolved seems valuable).
I don’t see anyone analogous to him, or even to someone who actually works at anthropic, the Enron witness list. Can you point to examples of analogous people in other cases?
As Tyler mentioned, it would be a massive effort to try to identify everyone who had been served with a discovery request or a deposition demand in a related civil proceeding. As in, it could take a hundred hours or more in Madoff (probably somewhat less in Enron), and the list might not be complete. So I don’t think that is a good use of anyone’s time.
I 100% agree that processing everyone who’d been touched by any Madoff case would be a massive undertaking and not worth the time. Which is why I didn’t ask for it. Right now I’m just asking for a single example of people in the reference class this post is discussing, being involved in the case. That’s definitely not the last question I’m going to ask, but it seems like a good starting place.
Maybe this is one of those dumb questions that’s so obvious to everyone with any expertise that they are laughing at me, and it takes a minute think of a specific example because it’s so common an expert would never bother to remember specifics. In that case, I expect experts to still be able to produce an example with a tiny amount of work once they stop laughing. And given legal advices’ reputation for overcaution in general, I think asking for examples and eventually numbers for a particular piece of advice is a fair request.
Alternately, I’d love to see legal experts replace Habryka’s guesstimate with more informed ones. My understanding from your comment here is that you’re not comfortable with that because the question isn’t well-defined. I think defining different scenarios and providing numbers for those (including some that are obviously materially different from this case, both to signal that the numbers can’t be applied directly and because it’s useful for teaching principles). Or add error bars until comfortable.
None of us are experts on the Madoff litigation itself—and even someone who had worked on it at a staff level would have only worked on a tiny sliver of it and would only have a small fraction of the universe of discovery targets in their mind.
I can partially quantify an estimate of how many entities experienced significant discovery burdens (e.g., sitting for a deposition, having to search and turn over significant documents, answering interrogatories) due to the Madoff civil litigation. Let’s set the lower error bar at 1,000 because there were 1,000 civil suits. Let’s guess a error bar extending to 10,000 based on Federal Rule of Civil Procedure 30(a)(2)(a)(i), which allows ten depositions per side without specific court approval—keeping in mind that there are other ways to burden a defendant or third party without taking their deposition. Although there is still significant uncertainty there, the number of civil suits (and the massive amount expended by the trustee) are far better predictors of someone getting tangled up in this than whether they were called to testify in a criminal trial.
I didn’t account for the fact that the Madoff trustee settled many of the estate’s claims without litigation—that means the 1,000 number is probably too low for our purposes. Nobody knows which claims will settle—which would affect which 1,000 to 10,000 get hit. So the universe of people who might reasonably expect to be in that 1,000 to 10,000 is maybe double that.
Are any EA leaders among the 20,000 people the debtor’s counsel would be most interested in hearing from? And could any statements by those people add them (or their colleagues/contacts) to that list, increase their exposures, or convince debtor’s counsel to spend more resources digging into them? That requires knowing more about the FTX situation, what the EA leaders know, and how their statements would read than any of us know.
We don’t know the facts—no one fully does at this stage—but there is an excellent chance the people who worked on the FTX Future Fund will be on a list, there is a strong chance that top officials at anyone who received more than $5MM from the Future Fund will be on a list (moderate chance at lower levels), there is a strong chance that anyone who was a regular correspondent with SBF or another key figure will be on a list. I believe most EA leaders are in one of those categories—or are an associate of someone who is in those categories and whose situation could be complicated by the first person speaking freely for reasons Tyler explained.
Thank you, this is exactly what I was hoping for! I have several questions.
Could you expand on how lawyers can create costs for 3rd parties, even without depositions?
I don’t understand why you’re using the exact same number as the Madoff clawbacks. I looked at it more for ratios: out of 40,000 investors (another source says 37k), 2,000 were net positive, and ~half of those were sued.* Your intuitions about where lines might be drawn were really helpful here, although I don’t feel like I have the level of clarity I want yet.
Suppose someone gets on a list: what happens next? how much of that is set in stone? What changes could someone induce by talking publicly? Some guesses I have are getting deposed versus not, deposition takes n% longer, personal records get subpoenaed where they otherwise wouldn’t have.
*Note that it’s not 1000+ cases that went to trial, it’s 1,000 lawsuits filed (apologies for leaving the citation out earlier, I meant to include it). I have yet to find any that went to trial, although some clearly got far enough to be prepping witnesses. That’s about what the trustee initially estimated- that out of 2000 net beneficiaries of Madoff, he’d go after about half of them. It’s not clear to me anyone gave back money without a lawsuit being filed.
As far as what kind of pain the legal process imposes on those who get caught up in it, the bulk comes prior to trial. So the fact that these cases didn’t go to trial doesn’t move the needle much. Actual trials in cases between midsize+ corporations are fairly uncommon, as the EV of the litigation is usually pretty clear after discovery ends and any summary-judgment motions are litigated to resolve disputed issues of law. Few cases of that sort are dependent on disputed witness testimony or the like.
The biggest practical pain point is document discovery, potentially having to go through potentially tens to hundreds of thousands of pages of documents to figure out which are responsive to the other side’s document demands and what has to be turned over. In a really simple case with 1,000 pages of material held by a few employees, responding to discovery can still potentially take several dozen hours of attorney time (plus client time), so let’s ballpark $10K on the low end per new third-party target. It is really hard to assess how much responsive material a third party could have. $100K might be a reasonable high-end estimate. For a major target like CEA/EV, maybe an order of magnitude higher.
For an entity that is already a target, let’s start with attorney fees. Although I expect entities on the other side spent less than the trustee’s $1.4B, half to 70 percent of that wouldn’t surprise me at all. So mid-six figures per target seems ballpark plausible on average. Much of that would be in discovery. Giving the opponent more information up front would likely allow them to justify more burdensome discovery to the judge. I would guestimate a 10-25% uplift in litigation costs per target from making a statement. Obviously it depends on what the statement says—but if the statement says nothing of substance, issuing it would have little value in the first place?
There’s also the risk of more easily triggering lawsuits from third party plaintiffs. Under Federal Rule of Civil Procedure 11, an attorney or party cannot file a complaint or pleading without conducting a reasonable investigation into the merits of the complaint (more or less). Giving a statement that isn’t 100 percent exculpatory makes it very easy for some would-be plaintiffs to meet their Rule 11 obligations and file a suit that wouldn’t have been filed otherwise. It’s possible, but less likely, that an organization’s statement could trigger a suit against a third party. Moreover, having a statement to hang their hat on likely makes it easier for would-be litigants to write their complaints in a way that survives the first, relatively inexpensive shot you get to get rid of a case (a Rule 12(b) motion to dismiss, see the Supreme Court’s decisions in Iqbal and Twombly if you want to geek out on that standard).
If you look at sanctions awards in frivolous cases (e.g., the “Kraken” election litigation in Michigan), you’ll see that defendants often can spend 250K on beating them back. That could be higher or lower based on the nature of the litigation. So I’m going to assign an EV of 250k as a rough estimate to a mid-to-larger org’s exposure for speaking out, which could represent a few really quick suits, one Kraken-level suit, or 0.1-0.2 of a more significant class action.
An organization also needs to be very careful in its public statements because they are admissions that can be used against it in court to devastating effect. That means it probably should conduct its own internal investigation before committing to a position that says anything substantive. Employees are known to . . . misremember or forget stuff when the facts don’t make them look good. Human nature. Depending on the organization’s risk tolerance, potential exposure, and amount of material, $25K-$250K per organization could be a reasonable estimate (but I’ve never been involved in pricing them, and I think a good investigation of CEA/EV would be higher given how much involvement it seems to have had with SBF directly).
Plus, a detailed statement gives the opponent a roadmap to make their discovery and litigation strategies more effective. At my upper-end estimate of the amount at stake, even a few percent increase in the estate’s chance of winning against EV/CEA alone could have a negative seven-figure EV. This would be lower for other organizations of course.
Cynically, there is also the chance that an organization knows it is toast, and that putting out a statement that would basically concede that is a great way to hasten the end. You can decide for yourself what the utility (or disutility) of keeping an organization running for an extra 2-3 years might be.
All of that is really conjectural on my end, but hopefully it gives you some basis for how a lawyer (who I emphasize has zero inside knowledge at all!) might try to quantify the costs of speaking rather than maintaining radio silence. I have low confidence in that quanitification. I would never attempt to quantify in this way for a client; the error bars are just too large.
Ultimately, it is a client decision whether the client should talk about pending litigation. Only in rare cases is silence not the correct decision from a legal perspective. However, clients make decisions against their legal interest for any number of reasons. Sometimes these are poor decisions, sometimes the client decides they have non-legal interests that they properly judge to outweigh their legal interests.
Apologies for the delay, and thank you, this is really helpful in fleshing out my models.
Am I correct that there is a ceiling on the amount of discovery that can be done? Once you have to turn over every document, there are no more documents to turn over. So the costs of speaking should asymptote over time? If not, what costs are increasingly linearly or faster?
> if the statement says nothing of substance, issuing it would have little value in the first place
is this true? I imagine there are some statements that are interesting to the EA community but bear no legal weight (“we believe fraud is bad”), and others contain information that would have been in discovery anyway.
I get the feeling one thing going on is that there are safe statements, but it’s hard for laypeople to identify them ahead of time (see: SBF outlining his incompetence as if that was a defense instead of evidence of a crime). So lawyers default to the much easier to follow “please shut the hell up”.
Do you have estimates for client time? I expect many EA orgs to care more about that than lawyers fees, especially if potential increases are only on the order of 25%.
There is likely a ceiling effect, either on the volume of discoverable things or the other side’s CEA of marginal discovery.
I think major orgs have made “we think fraud is bad” statements. I assume no one thinks fraud is good, or would admit to it if they did, so those statements have zero value to me.
Whether there are other safe statements depends on the organization’s exposure, but in many cases your lawyers would have to engage in a lot of independent factfinding to figure out if that’s the case (including needing to figure out information not held by your organization or publicly available). On the other hand, “no one here ever interacted with SBF, took FTX money, or gave him much thought” is likely safe if true and shouldn’t require much work.
There were no safe statements a lawyer could identify for SBF—that would require identifying, accessing, and ingesting all material info DOJ would obtain on him, which was/is impossible.
Really hard for me to estimate staff time given the nature of the case. You probably can substitute lawyer time for staff time for many tasks if you value staff time at mote than the lawyer’s rate.
If you’re doing a CEA, you should consider that discovery will probably happen under a protective order and any litigation will very likely be settled in a way where there is minimal leakage of the discovered info. So I don’t think it is correct to assume that unfavorable information will be publicly disclosed no matter what.
So I don’t think it is correct to assume that unfavorable information will be publicly disclosed no matter what.
This feels like a nonsequitur to me. The question is “is there information an org or person might like to disclose, but shouldn’t, due to its effect on legal risk (including procedural burden)?”. If the entity doesn’t want the information to be public then this question doesn’t arise (we could argue about whether orgs have a proactive moral duty to confess to crimes, but that’s definitely not the question here).
It sounds like you’re saying there is a ceiling on the burden for discovery, trial, etc, so at a certain point new statements can’t increase that burden, although they could have other negative consequences. Is that correct?
There is certainly a ceiling. Then again, your organization ceasing to exist because it’s overrun with legal bills and/or a crippling judgment is entered against it is a ceiling, and for smaller organizations (especially those that spent grant funds and are already going to get hit with a clawback) it may well be a realistic estimate of the ceiling.
As far as a non sequitur, I think we’re interested in somewhat different questions. My formulation is more: “What are the relative costs and benefits of making statements and releasing information on this matter?” I don’t find comparing benefits to only the increase in legal risk to be particularly decision-relevant on what organizations should do. I think almost any statement that is not “we’re giving back all the money, right now” is going to be net negative PR, and I think the increased risk of public disclosure of unfavorable information has to be considered to answer my question (it’s OK if it isn’t part of your question).
I’m not sure how to ask this but it seems like you’re...angry?… at me for trying to be precise about this? I’m not trying to challenge you, I just want to double crux to discover where our assessments differ. I thought that might have been any number of places you have expertise and I don’t, but now it seems like maybe you just don’t see value in public sense-making.
If the point of this is to try to quantify things like lawyer time, witness time, etc., in large matters like this, that’s going to be a REALLY heavy lift that the above just scratches the surface of and would require pulling together info from a huge number of sources, including public dockets that are daunting both in size and difficulty to work with. Not to knock the effort so far. It’s just a huge lift that would take hours and hours and require a lot of knowledge about how US court systems work to get right.
If one were to venture such a thing, a good place to start (but by no means a comprehensive source) would be fee applications in bankruptcy courts. That said, I’d advise against it. Feels like a lot of work that is very unlikely to move the needle in this discussion.
I agree the systemic effort is likely not worth it.
I think part of the problem is that, almost by definition, the people who have any expertise at all in law have also been inculcated in a culture that is extremely averse to both risks and quantified estimates. That leaves people without expertise but higher tolerance for risk and quantification (like Habryka) to come up with their own numbers and perhaps not respect the expertise behind the caution enough. I was hoping to unstick that log jam by introducing specifics that could be discussed at the object level. And I consider your suggestion to look at fee applications in bankruptcy court to be a success on that front, although the problem is very far from resolved.
Would you be willing to sit down with me and operationalize or parameterize some beliefs, which we can then send a legal researcher after? Or something like a double crux collaboration.
I find it completely plausible that there are a ton of hidden costs I could never find as an amateur, but I do want to have hard data instead of merely taking your word for it. I think using your guidance to figure out the right questions to get hard data on might be a huge win.
Habryka has confirmed he’s happy to pay for your time. Given that and the fact that he’s currently paying an expensive non-expert researcher (me), I’m extremely confident that if we find questions with the right level of tractability and meaning, he’ll cover the eventual legal researcher as well, although I didn’t wait for answer before posting this.
When you make a recommendation that people’s silence is based on “sound legal reasoning” you are implicitly claiming that you know the expected cost of the actions, and that the cost outweighs the benefits.
Yes, getting a robust estimate that is airtight is going to be hard, but in the OP you are already making a claim that you have an estimate of the relevant costs, and you should be able to quantify that estimate, at least with error bounds that could potentially span one or two orders of magnitude.
If you can’t actually put any expected cost to your estimate, or you are genuinely so uncertain about the actual cost here that you can’t give any number, then I don’t see how the reasoning in the OP checks out, since in that case it seems quite plausible that people are making a mistake by overestimating the cost, and the benefits actually hugely outweigh the costs.
Like, I understand that sometimes an estimate can be based on personal experience and intuitions whose generators are hard to elicit, but my sense is in this situation, if you are confident in the costs outweighing the benefits, then you should be able to produce some kind of estimate here (though I am not saying that this is a total universal, a basketball player can be confident they can make a shot, with very little ability to explain why they think that).
I would definitely be up for it! Happy to coordinate a good time via DM. Will also send you my phone number that you can call any time in the next few hours.
And even if someone did all that work, we wouldn’t know to what extent the quantum of legal effort/drama in a particular fraud-bankruptcy case was a good predictor of the quantum of legal effort/drama in this one. We would need to know significantly more about the facts than is publicly known to adjust the quantums from similar cases into a semi-reliable estimate for this one.
Very off-the-cuff here, so forgive any sloppiness. 2 and 3 seem like empirical questions about which I don’t have enough information to even hazard a guess. I’d at best be stoking a similar kind of confusion my original post attempts to warn about.
Re 1, if we’re talking in the abstract about criminal or civil liability, then I think the answer is probably yes (with potential exceptions for situations where after-the-fact actions might give rise to liability for things like perjury and witness tampering, etc.).
But if we’re talking about cost in time, treasure and mental well-being associated with being a witness, or the knock-on effects for others and the system more generally, I think the answer is a strong no. Loose talk after a dispute has arisen can easily take a near-0% chance of someone being significantly involved (looking exclusively at pre-dispute facts) and make it a near certainty. And this would not necessarily be just for the individual or organization doing the talking. A simple but hopefully clear example would be where a central figure in the pre-dispute facts speaks post-dispute and identifies someone else as being significantly involved. The identified person would suddenly be on everyone’s radar even if they were on nobody’s radar before the statement, and that would be true even they weren’t actually involved (i.e., the post-dispute statement was false), and even if the false statement was an innocent mistake (a few examples of which I try to outline in the post—e.g., confusing post-dispute involvement with pre-dispute involvement, etc.).
If I understand you correctly, you’d agree that carefully saying true things, with extra attention to clarity (to avoid easy misreadings), would go a long way to reducing risk of (1).
I’m not sure I understand this comment completely, but I think I disagree with it. Speaking in the abstract because I don’t want to suggest I know anything about any specifics here, I think the best way to avoid legal risk (including but not limited to liability) based on pre-blow-up actions or statements in any situation like this is to stay silent after the fact and only say what is absolutely necessary through the formal legal process.
I am not Tyler but am a lawyer. Based on limited information, I think it likely that over 90% of legal exposure would be based on actions that happened on or before November 11 (let’s use the date of the bankruptcy filing). However, that is not equivalent to saying that what potential targets do from November 11 onward will affect the ultimate financial pain they end up bearing (damages, clawbacks, legal fees, intangible negatives) by no more than 10%. I think the latter statement is likely false, and is the question one has to consider when deciding on speech vs. silence.
The answer to #2 depends in part of your definitions of “prominent EA figures” or “involved in the litigation process.”
I would think less of Sullivan & Cromwell’s litigation team if EV and CEA are not involved in litigation in some capacity. I have zero inside knowledge, but way too much money went from FTX-affiliated entities and persons to EV and CEA not to make an attempt to clawback and to see what else might be there.
[Edited: “on or before November 11,” “November 11 onward.” Thanks to the person who messaged me to note those typos!]
I want to push back on this post. I think sadly this post suffers from the same problem that 99% of all legal advice that people receive suffers from, which is that it is not actually a risk analysis that helps you understand the actual costs of different decisions.
The central paragraph that I think most people will react to is this section:
Sadly, this post does not indicate what the actual expected time a witness might be expected to spend in litigation is.
I don’t currently know this number, but my guess is that less than 10% of any EA leaders who would end up writing publicly about this would end up being involved in any substantial amount of legal proceedings. My guess based on doing some independent legal research myself is that the actual cost of being involved in this kind of litigation is substantial, but not massive. Most likely a few thousand dollars in lawyer fees (for the simplicity of calculation, let’s say $10k), and probably a few days of being involved in court proceedings (let’s say for simplicity 5 days, spending like 5 hours on each of those days).
Finishing this very rough estimate, my guess is the actual expected cost of writing publicly about this is around 0.1 * $10k + 0.1 * 25 hours = $1k + 2.5 hours. My guess is this is well below what most EA leaders would be willing to pay to be able to speak openly and publicly about this, and would at most be a minor consideration on whether to actually say anything on this topic.
That said, I am not a legal professional, and my estimates might be totally wrong here. I would really appreciate a legal analysis that actually estimates the cost here, since I think as written, it seems totally plausible to me (and indeed more likely than not) that the expected cost of speaking up is pretty minor, though it is also plausible that the expected cost could be major, and settling which one it is is the primary question that legal analysis should aim to answer, in my opinion.
For related analysis that captures my feelings on this topic, see also Everyone Should be Less Afraid of Lawsuits by Alyssa Vance, in-particular this section:
I consider your attempt at a quantified expected cost analysis a helping hand, not pushback, and I appreciate it.
Accepting it as a data point, a few quick points in response:
Your comment only addresses the self-interest angle, which was a relatively small part of my post. It (understandably) ignores the impacts on others and the systemic impacts that I tried to highlight, which I don’t think can be disentangled from the self-interest analysis so easily. I’m not sure those additional impacts are amenable to quantified expectation analysis (though I’d be happy to be proven wrong on that), but we shouldn’t just ignore them.
I think your numbers are low at the outset, but I don’t think any tweaking I’d do would cause us to be off by an OOM. That said, I think you’ve established a floor, and one that only applies to a witness with no potential liability. Accepting your numbers for the sake of discussion, the time estimates sit at the bottom of towering error bars. And that’s assuming we’re talking about an individual and not an organization that might have orders of magnitude more documents to review than an individual would, attorney-client privilege and other concerns that complicate things greatly and require a great deal more attorney time, and potentially many individuals involved as witnesses, with all the associated costs for each one.
The expectation analysis also (again, understandably) only addresses time and money and does not address mental and emotional impacts. Such things are obviously difficult to quantify, but they are worth taking seriously. The legal system fails to address these issues at basically every turn, and causes a great deal of suffering as a result.
On the point about lawyers being overly conservative, I think from a 10,000-foot view assessing emergent trends across all legal advice given in all situations, I agree with this. But when you dig into individual instances of legal advice, the idiosyncrasies of each situation make applying a heuristic based on emergent macro-trends difficult to justify. I recognize that’s not a solution to what I agree with you is a problem, but I think it’s an important part of the problem to recognize in the context of this discussion.
Appreciate the thoughtful response.
Thank you! I am glad.
Some further thoughts in response:
I do admit to failing to understand your case that the impact on others and the systemic impacts are expected to be negative, not positive.
It seems that our prior should be strongly that making the courts have more relevant information will overall cause them to make better decisions, and while you list some plausible scenarios of negative impact, the scenario of “someone says something true publicly that then gets used in the lawsuit to help identify the circumstances and details of the crimes that appear to have been committed in a way that both makes justice more likely to be achieved and saves everyone time” seems by far the more likely outcome.
My guess is that if you had a class action lawsuit by the FTX credits or people who deposited their money into FTX that they would pretty strongly prefer more people go give more information on FTX publicly, since that will likely make their job easier, not harder, and as such I think the sign of the secondary effects here points strongly in a positive, not negative, direction.
I agree that if you have potential liability the costs might be higher, though it’s a bit unclear what our moral imperative in that case should be (like, depends on whether your liability here is the result of kind of crazy laws or not).
I think there are definitely some individuals who have more at stake here, and where it probably makes sense to be more quiet, though I would be quite surprised if this includes most of the EA leadership.
I feel a bit confused about how much organizations could get involved in this situation, and would appreciate more datapoints on what it means for an organization to get involved in this at a level higher than individual witnesses. I understand the obligations that individuals are under if they are called in for testimony, but don’t understand the obligations that organizations could be under in a case like this.
I agree that there are substantial costs here, but I also want to talk about the symmetric costs on the other side that strike me as substantially greater. My sense is most people will suffer a good amount more stress from trying to be quiet about this topic for many years, and will incur much greater costs from there being a lot of latent concern and anxiety for many years as people stay quiet, and that the mental and emotional impacts of that will be much greater than the impact of being involved in the proceedings itself, at least in the aggregate (I do think some individuals will have a pretty bad time with the court system, but I think approximately everyone in EA will have a pretty bad time if discussion around this topic stays guarded for many years).
Criticism of the US model of a strongly “adversarial” justice system is definitely valid; it wastes a lot of resources on inefficiently resolving certain sorts of disputes. However, the bankruptcy case is in a US federal court, and at least most of the key players are within the jurisdiction of said court. And I don’t think you usually get to a fair, equitable, or efficient outcome by not asserting your rights and interests when the other side is going out with guns blazing. The system tends to give both sides some good cards and some cards that suck; getting an outcome that is fair to all involved generally involves giving up your good cards in exchange for the other side giving up theirs too.
There is a parallel system of alternative dispute resolution (ADR) like mediation which tends to be much more efficient and more likely to produce outcomes everyone can live with. But it requires give and take by both sides, and right now the debtors are not in a position to engage with it. They have much more pressing issues to deal with in the next few months, like trying to track down $10B in missing crypto before it ends up untraceable or outside the practical control of a US court.
In my view, many organizations do have a moral obligation to return some of the funds—and an ADR process is the best way to do that. It is reasonable for the organizations to ask for dismissal of any claimed liabilities they should not morally have to bear in exchange for that.
Most litigation of note will be against organizations—like Willie Sutton didn’t actually say about banks, because that’s where the money is. To me, the likely story is that corporate counsel is advising organizations about the desirability of directing their employees to stay mum to protect the organization’s legal interests. In fact, organizational counsel does not represent the employees (which is a good thing to remember if you ever are involved with a lawyer at work!) That directive may incidentally benefit the employees and directors, though.
It is really difficult to give specific examples of the corporate/individual interplay because we don’t know what the identified legal risks are. But two general principles: first, an employer is generally liable for an employee’s actions in the scope of the employee’s employment. What “in the scope” means is pretty fact-specific. As an extreme example, we have the Department of Justice arguing at the moment that denying allegations of sexual misconduct from decades ago is in the scope of employment of the (former) U.S. president.
Second, corporations lack minds and cannot “know” anything other than through affiliated people. So for instance, if the CEO of Organization X comes on here and writes that he knew A, B, and C (at a time they were CEO), that will probably establish that Organization X knew A, B, and C. That might or might not be the case if a more junior employee said that they knew A, B, and C but higher-ups did not know. All that is to say that the speech of affiliated individuals tends to have effects on the legal interests of the organization too.
Unfortunately, because none of us know what is being discussed by the people involved, we are limited to speculation about what might be going on. Speculation is a poor substitute for information, but it can’t leak knowledge (that the speculator does not have) or be attributed to a person or organization (with whom the speculator has no affiliation). I know that is not very satisfying.
Sorry for the delay in responding here, and apologies in advance for keeping this one short.
Your position appears to assume that speaking outside the court process after the fact is likely to add accurate information to the public record and help everyone get to the truth. One of the main thrusts of my OP is that this is not the case. Instead, speaking outside the process after the fact is more likely to create confusion instead of clarity, along with all the other negative effects I walk through. I’m speaking strictly from my experience and in the abstract here, but I think that’s where we disagree. Even information provided carefully, with counsel, under the controlled circumstances of the judicial process can be problematic, but it at least has the benefit of being relatively structured and under oath. It’s much more likely to help everyone (including class action attorneys) get as close to the truth as possible.
I’m not saying there are not potential upsides of speaking out—including to the mental and emotional well-being of folks who want to speak out and have been advised to stay quiet. I’m just stating my view that those upsides are outweighed by the potential downsides, not only to the speaker, but to folks they might speak about and the process in general.
Re the question about how organizations could get involved, I don’t pretend to know the specifics here and wouldn’t comment if I thought I did, but on a general level Jason covers this in his response.
I appreciate your take here, but would also take bets at pretty large odds that if we were to ask the relevant creditors and defendants that they would prefer people to speak openly here, and that this guess of yours is inaccurate.
I think I just don’t understand the basic argument by which information here will harm people, especially if that information provides accurate evidence about actual wrongdoing that happened, or provides substantial evidence about who might be relevant witnesses to the case, by modeling the social dynamics around FTX.
The prior in favor of “accurate information helps people come to more accurate beliefs” just seems so much stronger here than the relatively vague and abstract arguments you are providing.
A simple model, which is roughly what I interpret legal experts to mean when they say that the US justice system is intentionally designed as adversarial, is something like: the probability that a judge finds you guilty (pj) is the product of the true probability that you are guilty (pt), the bias on evidence introduced by the prosecutor (bp), and the bias on evidence introduced by the defender (bd):[1]
pj=ptbpbd
We observe that pj=pt↔bp=b−1d i.e. the optimal truth finding strategy is when both sides are equally biased in their own favor. And importantly to your point, one side being unbiased results in the judge making a worse decision (assuming the other side is biased).
Furthermore, the equilibrium of this game is when sides are maximally biased in their own favor (since otherwise the more biased side would disproportionately win). Knowing this, the US legal system encourages both sides to be maximally biased, since this results in the most accurate judgments.
I would be interested to hear from others if this seems accurate, as it’s not something I know a lot about.
After writing this I realized these should be odds ratios, not probabilities, but hopefully you get the point.
FWIW, I’d happily take the other side of a bet from you here, if one could construct a wager that would settle this. But I don’t think that’s possible, which I reckon is part of the problem.
And while I’ll confess to being a little miffed at having my own prior (updated over 15+ years directly observing analogous circumstances) downgraded to a ”guess” in your comment, while your own remains on its linguistic perch, I don’t think our continuing to go around on this will be useful. I’ll propose we “settle” it by my buying you a beer (or a cup of coffee, if you prefer) the next time we’re in the same city (and I’ll draw my home circle out to include NYC, if you happen to come through), and we can spend the time discussing the many things I’d wager we agree on.
For the non-lawyers, I’d add that two of the factors you may not be thinking about are (A) the cost of the attorney gathering enough information to make even a semi-reliable estimate of expected costs; and (B) the risk of client error from various sources.
As a general matter: Client officials are not lawyers and may not understand what information is relevant to bring to the lawyer, even with some guidance they will need to err on the side of overproduction. So the lawyer often needs to review a lot of irrelevant information to better ascertain the actual risks involved, which raises costs.
Second, client officials have their own motivations which may run at cross purposes to the lawyer making an accurate risk assessment. Let’s say—hypothetically—that an employee did something that is in retrospect really inadvisible and that exposed the organization (and possibly themselves) to significant liability. That the attorney represents the corporation and not the employee, and should be explaining that to employees if there is any risk of a conflict of interest between the employee and the corporation. How likely is it that the employee is going to ’fess up with minimal prompting?
I think you’re quite right to be asking these questions (although I think your estimate of expected cost is significantly too low for many leaders and organizations).
A good lawyer (whose client is willing and able to pay for good advice) generally does not issue commands. They take time to understand the client’s interests, objectives, and tradeoffs and then provide professional advice explaining risks/tradeoffs to guide the client as to how to move forward. The choice of objectives belongs to the client as long as those objectives are lawful. Unfortunately, the client can’t tell the rest of the community about what risks were considered without blowing the attorney-client privilege and providing very valuable information to the other side.
The original comment is probably vague on the risk/benefit tradeoff in part because Tyler is not any organization’s lawyer and does not have the information needed to assess each organization’s risk specifically. I took his main point to be that non-lawyers would be prone to underestimate the various costs of speaking out, and that it wouldn’t be appropriate to draw any inferences from anyone’s decision not to do so. I don’t think he was trying to provide legal advice on whether anyone should speak out. The benefits of speaking out are non-legal benefits that are really for the client to weigh given its objectives, and have been discussed elsewhere at some length on this forum.
At Oliver’s request I dug into the Madoff case a little, to see if there was anything we could learn from.
Epistemic status: I am not a lawyer and I spent 2.5 hours on this, one of which was unproductively wrestling with the court records database, the worst database UI in the history of databases.
The following are court cases I found related to Madoff Investments:
Bernie Madoff pled guilty and never went to trial. It wasn’t even a negotiated plea bargain, just a straight guilty please (probably so he didn’t have to testify against anyone)
His two top lieutenants took plea deals and never went to trial (but did testify elsewhere).
5 other employees had a joint criminal trial. The Madoff case broke in 2008, they were indicted in 2010 (not all together), the case went to trial in 2013, and took 8 months to complete. Appeals continued until 2016.
The main trustee in charge of clawback filed “over 1,000” civil suits aimed at clawing back money. None of the ones I found in the news went to trial, although some got very close. I probably identified <20 cases, primarily the largest ones.
The lawsuits were (reportedly) aimed at people who took more money out than they had ever put in. The ones that made the news typically had a component of “knew or should have known Madoff was a fraud” and encouraging others to invest. However the money recovered went to a general fund, not the people they specifically were responsible for bringing to Madoff.
I found a handful of other lawsuits brought against the same entities from the large set of civil suits. None of these appear to have gone to trial, although they got very little coverage so it’s hard to say.
So that’s one trial. I spent an hour trying to get a witness list and failing, and am now hoping someone reads this and Cunningham’s laws me. You’re looking for S.D.N.Y., No. 09-mag-2484 or 1:10-cr-00228, transcripts are $6+ a page to get quickly. You might be able to do better going to the courthouse in NYC.
In the meantime:
That trial had “over 40” witnesses. Let’s call it 43.
The trial lasted 8 months, although not ever day had witness testimony.
Of those witnesses I found the following:
2 Expert witnesses/investigators who clearly opted in
Matthew Cohen, Bruce Dubinsky
6Employees of Madoff who could predict they would be called in (this includes the plea bargainers)
Winnifer Jackson, David Kugel, Enrica Cotellessa-Pitz, Friehling, Konigsburg, DiPascali
2 defendents
Bonventre, Bongiorno
That leaves 30+ witnesses unaccounted for, which isn’t really enough to draw interesting conclusions, especially without knowing the denominator of who could have been called.
The main Enron trial was much more cooperative and provided full lists of witnesses, with profiles.
Wall Street Journal profiles of prosecution witnesses:
Enron Employees and contractors: 20
Arthur Andersen Employes: 2
Criminal Investigators: 1
Other: 2
Rob Barone, managing director at S&P who previously oversaw an assessment of Enron
Glenn Ray, a stockbroker who managed a defendent’s sale of Enron stock
Wall Street Journal profiles of defense witnesses:
Enron employees: 15
11 material
3 character witnesses (which is presumably voluntary)
1 contractor
Defendants: 2
Expert Witnesses: 3
Non-enron character witnesses: 11
The prosecution also shared its prospective witness list. It has 31 names, of which 23 testified, plus another 2 who weren’t on the original list. Of those two, one was an FBI agent for whom this was part of his job, and another was an Enron employee brought in as a rebuttal witness.
Enron produced a lot of trials, of which this is only one. There were so many trials Wikipedia didn’t even bother to list all of them (example), plus the senate hearings, extradition trials, appeals, etc. This is a very narrow slice of who had to testify about Enron.
But all of the witnesses either opted in (as character or expert witnesses, or investigators), worked for the company or person that committed crimes, or worked for the auditing firm that missed the crimes. None of these seem equivalent to the people I imagine you are talking to or about here. Could you give some specific examples of witnesses in positions similar to those people, in cases similar to FTX’s likely case? I’m happy with a very broad definition of similar, and I would absolutely believe they exist and I missed them. I’m a non-expert who has looked at two cases so far and both are missing crucial aspects of EA’s relationship to FTX. But I’m clearly not going to find them very fast.
Full disclosure: habryka is paying me for this research. He didn’t review either comment before publication but I did talk to him about my findings as I was working. I meant to include this in the first comment but the “(paid)” in “At Oliver’s (paid) request” got lost in an edit somewhere.
I would maybe edit it in?
I’m not sure what conclusion this would support.
The scope of what the prosecution has to prove in a criminal prosecution can be surprisingly narrow. Specifically, the prosecution needs to prove the elements of the charged offenses—not the entire scope of the criminal enterprise. Assuming there is a conviction, other information can come in at sentencing where the proof is much simpler. In contrast, in civil litigation, you need a much more complete picture to establish who the bankruptcy estate might be able to get money from.
As this article notes, the Madoff trustee’s firm ran up $1.4B in legal fees as of last year on civil litigation. You note that they filed about 1,000 civil lawsuits—I suspect there were significantly more targets than that because the trustee viewed litigation as a last resort when settlement negotiations failed. I have no way of knowing how many different individuals and entities got hit with depositions and discovery demands in those 1,000+ cases . . . but I would suspect the average number of depositions and discovery demands in a filed civil case is greater than one each. The trustee had to be doing something substantial to justify the court approving all those fees . . .
I suspect there was less civil litigation around Enron, but for reasons that aren’t helpful here. Enron was a publicly-traded corporation, and IIRC the bulk of the losses were equity (people who held Enron stock that was overvalued due to fraudulent financial reports). Neither Enron nor those who lost money on Enron stock would be expected to have a claim against people who traded in its stock without knowledge of the fraud. Among other reasons, Enron itself wasn’t a party to those transactions, so its bankruptcy estate would have no clawback claims against third parties who sold Enron stock to other third parties for more than it was ultimately worth.
I don’t think this supports any conclusions yet, it’s two cases and I only got a full witness list for one of those. I would love to see any examples of witnesses who were in roughly the position EA leadership or members (allegedly) are- people who received money for charitable purposes, from someone who earned it through fraud and needs to pay it back, but weren’t close relatives or could otherwise be considered tightly aligned with their funder. Or someone like Holden, whose org AFAIK didn’t receive any FTX money at all.
(FTR I think Holden waiting until his takes are cold is entirely reasonable, I think the speed of response on the EAF is overheated, but given the trials could take over a decade, establishing the risks of speaking candidly before they are entirely resolved seems valuable).
Doesn’t Holden have close personal ties to Anthropic, which is potentially a huge target?
Oh I had indeed forgotten that.
I don’t see anyone analogous to him, or even to someone who actually works at anthropic, the Enron witness list. Can you point to examples of analogous people in other cases?
As Tyler mentioned, it would be a massive effort to try to identify everyone who had been served with a discovery request or a deposition demand in a related civil proceeding. As in, it could take a hundred hours or more in Madoff (probably somewhat less in Enron), and the list might not be complete. So I don’t think that is a good use of anyone’s time.
I 100% agree that processing everyone who’d been touched by any Madoff case would be a massive undertaking and not worth the time. Which is why I didn’t ask for it. Right now I’m just asking for a single example of people in the reference class this post is discussing, being involved in the case. That’s definitely not the last question I’m going to ask, but it seems like a good starting place.
Maybe this is one of those dumb questions that’s so obvious to everyone with any expertise that they are laughing at me, and it takes a minute think of a specific example because it’s so common an expert would never bother to remember specifics. In that case, I expect experts to still be able to produce an example with a tiny amount of work once they stop laughing. And given legal advices’ reputation for overcaution in general, I think asking for examples and eventually numbers for a particular piece of advice is a fair request.
Alternately, I’d love to see legal experts replace Habryka’s guesstimate with more informed ones. My understanding from your comment here is that you’re not comfortable with that because the question isn’t well-defined. I think defining different scenarios and providing numbers for those (including some that are obviously materially different from this case, both to signal that the numbers can’t be applied directly and because it’s useful for teaching principles). Or add error bars until comfortable.
None of us are experts on the Madoff litigation itself—and even someone who had worked on it at a staff level would have only worked on a tiny sliver of it and would only have a small fraction of the universe of discovery targets in their mind.
I can partially quantify an estimate of how many entities experienced significant discovery burdens (e.g., sitting for a deposition, having to search and turn over significant documents, answering interrogatories) due to the Madoff civil litigation. Let’s set the lower error bar at 1,000 because there were 1,000 civil suits. Let’s guess a error bar extending to 10,000 based on Federal Rule of Civil Procedure 30(a)(2)(a)(i), which allows ten depositions per side without specific court approval—keeping in mind that there are other ways to burden a defendant or third party without taking their deposition. Although there is still significant uncertainty there, the number of civil suits (and the massive amount expended by the trustee) are far better predictors of someone getting tangled up in this than whether they were called to testify in a criminal trial.
I didn’t account for the fact that the Madoff trustee settled many of the estate’s claims without litigation—that means the 1,000 number is probably too low for our purposes. Nobody knows which claims will settle—which would affect which 1,000 to 10,000 get hit. So the universe of people who might reasonably expect to be in that 1,000 to 10,000 is maybe double that.
Are any EA leaders among the 20,000 people the debtor’s counsel would be most interested in hearing from? And could any statements by those people add them (or their colleagues/contacts) to that list, increase their exposures, or convince debtor’s counsel to spend more resources digging into them? That requires knowing more about the FTX situation, what the EA leaders know, and how their statements would read than any of us know.
We don’t know the facts—no one fully does at this stage—but there is an excellent chance the people who worked on the FTX Future Fund will be on a list, there is a strong chance that top officials at anyone who received more than $5MM from the Future Fund will be on a list (moderate chance at lower levels), there is a strong chance that anyone who was a regular correspondent with SBF or another key figure will be on a list. I believe most EA leaders are in one of those categories—or are an associate of someone who is in those categories and whose situation could be complicated by the first person speaking freely for reasons Tyler explained.
Thank you, this is exactly what I was hoping for! I have several questions.
Could you expand on how lawyers can create costs for 3rd parties, even without depositions?
I don’t understand why you’re using the exact same number as the Madoff clawbacks. I looked at it more for ratios: out of 40,000 investors (another source says 37k), 2,000 were net positive, and ~half of those were sued.* Your intuitions about where lines might be drawn were really helpful here, although I don’t feel like I have the level of clarity I want yet.
Suppose someone gets on a list: what happens next? how much of that is set in stone? What changes could someone induce by talking publicly? Some guesses I have are getting deposed versus not, deposition takes n% longer, personal records get subpoenaed where they otherwise wouldn’t have.
*Note that it’s not 1000+ cases that went to trial, it’s 1,000 lawsuits filed (apologies for leaving the citation out earlier, I meant to include it). I have yet to find any that went to trial, although some clearly got far enough to be prepping witnesses. That’s about what the trustee initially estimated- that out of 2000 net beneficiaries of Madoff, he’d go after about half of them. It’s not clear to me anyone gave back money without a lawsuit being filed.
As far as what kind of pain the legal process imposes on those who get caught up in it, the bulk comes prior to trial. So the fact that these cases didn’t go to trial doesn’t move the needle much. Actual trials in cases between midsize+ corporations are fairly uncommon, as the EV of the litigation is usually pretty clear after discovery ends and any summary-judgment motions are litigated to resolve disputed issues of law. Few cases of that sort are dependent on disputed witness testimony or the like.
The biggest practical pain point is document discovery, potentially having to go through potentially tens to hundreds of thousands of pages of documents to figure out which are responsive to the other side’s document demands and what has to be turned over. In a really simple case with 1,000 pages of material held by a few employees, responding to discovery can still potentially take several dozen hours of attorney time (plus client time), so let’s ballpark $10K on the low end per new third-party target. It is really hard to assess how much responsive material a third party could have. $100K might be a reasonable high-end estimate. For a major target like CEA/EV, maybe an order of magnitude higher.
For an entity that is already a target, let’s start with attorney fees. Although I expect entities on the other side spent less than the trustee’s $1.4B, half to 70 percent of that wouldn’t surprise me at all. So mid-six figures per target seems ballpark plausible on average. Much of that would be in discovery. Giving the opponent more information up front would likely allow them to justify more burdensome discovery to the judge. I would guestimate a 10-25% uplift in litigation costs per target from making a statement. Obviously it depends on what the statement says—but if the statement says nothing of substance, issuing it would have little value in the first place?
There’s also the risk of more easily triggering lawsuits from third party plaintiffs. Under Federal Rule of Civil Procedure 11, an attorney or party cannot file a complaint or pleading without conducting a reasonable investigation into the merits of the complaint (more or less). Giving a statement that isn’t 100 percent exculpatory makes it very easy for some would-be plaintiffs to meet their Rule 11 obligations and file a suit that wouldn’t have been filed otherwise. It’s possible, but less likely, that an organization’s statement could trigger a suit against a third party. Moreover, having a statement to hang their hat on likely makes it easier for would-be litigants to write their complaints in a way that survives the first, relatively inexpensive shot you get to get rid of a case (a Rule 12(b) motion to dismiss, see the Supreme Court’s decisions in Iqbal and Twombly if you want to geek out on that standard).
If you look at sanctions awards in frivolous cases (e.g., the “Kraken” election litigation in Michigan), you’ll see that defendants often can spend 250K on beating them back. That could be higher or lower based on the nature of the litigation. So I’m going to assign an EV of 250k as a rough estimate to a mid-to-larger org’s exposure for speaking out, which could represent a few really quick suits, one Kraken-level suit, or 0.1-0.2 of a more significant class action.
An organization also needs to be very careful in its public statements because they are admissions that can be used against it in court to devastating effect. That means it probably should conduct its own internal investigation before committing to a position that says anything substantive. Employees are known to . . . misremember or forget stuff when the facts don’t make them look good. Human nature. Depending on the organization’s risk tolerance, potential exposure, and amount of material, $25K-$250K per organization could be a reasonable estimate (but I’ve never been involved in pricing them, and I think a good investigation of CEA/EV would be higher given how much involvement it seems to have had with SBF directly).
Plus, a detailed statement gives the opponent a roadmap to make their discovery and litigation strategies more effective. At my upper-end estimate of the amount at stake, even a few percent increase in the estate’s chance of winning against EV/CEA alone could have a negative seven-figure EV. This would be lower for other organizations of course.
Cynically, there is also the chance that an organization knows it is toast, and that putting out a statement that would basically concede that is a great way to hasten the end. You can decide for yourself what the utility (or disutility) of keeping an organization running for an extra 2-3 years might be.
All of that is really conjectural on my end, but hopefully it gives you some basis for how a lawyer (who I emphasize has zero inside knowledge at all!) might try to quantify the costs of speaking rather than maintaining radio silence. I have low confidence in that quanitification. I would never attempt to quantify in this way for a client; the error bars are just too large.
Ultimately, it is a client decision whether the client should talk about pending litigation. Only in rare cases is silence not the correct decision from a legal perspective. However, clients make decisions against their legal interest for any number of reasons. Sometimes these are poor decisions, sometimes the client decides they have non-legal interests that they properly judge to outweigh their legal interests.
Apologies for the delay, and thank you, this is really helpful in fleshing out my models.
Am I correct that there is a ceiling on the amount of discovery that can be done? Once you have to turn over every document, there are no more documents to turn over. So the costs of speaking should asymptote over time? If not, what costs are increasingly linearly or faster?
> if the statement says nothing of substance, issuing it would have little value in the first place
is this true? I imagine there are some statements that are interesting to the EA community but bear no legal weight (“we believe fraud is bad”), and others contain information that would have been in discovery anyway.
I get the feeling one thing going on is that there are safe statements, but it’s hard for laypeople to identify them ahead of time (see: SBF outlining his incompetence as if that was a defense instead of evidence of a crime). So lawyers default to the much easier to follow “please shut the hell up”.
Do you have estimates for client time? I expect many EA orgs to care more about that than lawyers fees, especially if potential increases are only on the order of 25%.
There is likely a ceiling effect, either on the volume of discoverable things or the other side’s CEA of marginal discovery.
I think major orgs have made “we think fraud is bad” statements. I assume no one thinks fraud is good, or would admit to it if they did, so those statements have zero value to me.
Whether there are other safe statements depends on the organization’s exposure, but in many cases your lawyers would have to engage in a lot of independent factfinding to figure out if that’s the case (including needing to figure out information not held by your organization or publicly available). On the other hand, “no one here ever interacted with SBF, took FTX money, or gave him much thought” is likely safe if true and shouldn’t require much work.
There were no safe statements a lawyer could identify for SBF—that would require identifying, accessing, and ingesting all material info DOJ would obtain on him, which was/is impossible.
Really hard for me to estimate staff time given the nature of the case. You probably can substitute lawyer time for staff time for many tasks if you value staff time at mote than the lawyer’s rate.
If you’re doing a CEA, you should consider that discovery will probably happen under a protective order and any litigation will very likely be settled in a way where there is minimal leakage of the discovered info. So I don’t think it is correct to assume that unfavorable information will be publicly disclosed no matter what.
This feels like a nonsequitur to me. The question is “is there information an org or person might like to disclose, but shouldn’t, due to its effect on legal risk (including procedural burden)?”. If the entity doesn’t want the information to be public then this question doesn’t arise (we could argue about whether orgs have a proactive moral duty to confess to crimes, but that’s definitely not the question here).
It sounds like you’re saying there is a ceiling on the burden for discovery, trial, etc, so at a certain point new statements can’t increase that burden, although they could have other negative consequences. Is that correct?
There is certainly a ceiling. Then again, your organization ceasing to exist because it’s overrun with legal bills and/or a crippling judgment is entered against it is a ceiling, and for smaller organizations (especially those that spent grant funds and are already going to get hit with a clawback) it may well be a realistic estimate of the ceiling.
As far as a non sequitur, I think we’re interested in somewhat different questions. My formulation is more: “What are the relative costs and benefits of making statements and releasing information on this matter?” I don’t find comparing benefits to only the increase in legal risk to be particularly decision-relevant on what organizations should do. I think almost any statement that is not “we’re giving back all the money, right now” is going to be net negative PR, and I think the increased risk of public disclosure of unfavorable information has to be considered to answer my question (it’s OK if it isn’t part of your question).
I would note that there is a huge jump from “unfavorable information” to “confess to crimes”—for instance, there is generally no duty to rescue in common-law jurisdictions (which would include the US and UK).
I’m not sure how to ask this but it seems like you’re...angry?… at me for trying to be precise about this? I’m not trying to challenge you, I just want to double crux to discover where our assessments differ. I thought that might have been any number of places you have expertise and I don’t, but now it seems like maybe you just don’t see value in public sense-making.
My two cents: I don’t read any negative emotions from Jason’s comment
If the point of this is to try to quantify things like lawyer time, witness time, etc., in large matters like this, that’s going to be a REALLY heavy lift that the above just scratches the surface of and would require pulling together info from a huge number of sources, including public dockets that are daunting both in size and difficulty to work with. Not to knock the effort so far. It’s just a huge lift that would take hours and hours and require a lot of knowledge about how US court systems work to get right.
If one were to venture such a thing, a good place to start (but by no means a comprehensive source) would be fee applications in bankruptcy courts. That said, I’d advise against it. Feels like a lot of work that is very unlikely to move the needle in this discussion.
I agree the systemic effort is likely not worth it.
I think part of the problem is that, almost by definition, the people who have any expertise at all in law have also been inculcated in a culture that is extremely averse to both risks and quantified estimates. That leaves people without expertise but higher tolerance for risk and quantification (like Habryka) to come up with their own numbers and perhaps not respect the expertise behind the caution enough. I was hoping to unstick that log jam by introducing specifics that could be discussed at the object level. And I consider your suggestion to look at fee applications in bankruptcy court to be a success on that front, although the problem is very far from resolved.
Would you be willing to sit down with me and operationalize or parameterize some beliefs, which we can then send a legal researcher after? Or something like a double crux collaboration.
I find it completely plausible that there are a ton of hidden costs I could never find as an amateur, but I do want to have hard data instead of merely taking your word for it. I think using your guidance to figure out the right questions to get hard data on might be a huge win.
Habryka has confirmed he’s happy to pay for your time. Given that and the fact that he’s currently paying an expensive non-expert researcher (me), I’m extremely confident that if we find questions with the right level of tractability and meaning, he’ll cover the eventual legal researcher as well, although I didn’t wait for answer before posting this.
When you make a recommendation that people’s silence is based on “sound legal reasoning” you are implicitly claiming that you know the expected cost of the actions, and that the cost outweighs the benefits.
Yes, getting a robust estimate that is airtight is going to be hard, but in the OP you are already making a claim that you have an estimate of the relevant costs, and you should be able to quantify that estimate, at least with error bounds that could potentially span one or two orders of magnitude.
If you can’t actually put any expected cost to your estimate, or you are genuinely so uncertain about the actual cost here that you can’t give any number, then I don’t see how the reasoning in the OP checks out, since in that case it seems quite plausible that people are making a mistake by overestimating the cost, and the benefits actually hugely outweigh the costs.
Like, I understand that sometimes an estimate can be based on personal experience and intuitions whose generators are hard to elicit, but my sense is in this situation, if you are confident in the costs outweighing the benefits, then you should be able to produce some kind of estimate here (though I am not saying that this is a total universal, a basketball player can be confident they can make a shot, with very little ability to explain why they think that).
Does it make sense to have a quick chat about this in realtime? I’d be open to that and think it would be more productive.
I would definitely be up for it! Happy to coordinate a good time via DM. Will also send you my phone number that you can call any time in the next few hours.
And even if someone did all that work, we wouldn’t know to what extent the quantum of legal effort/drama in a particular fraud-bankruptcy case was a good predictor of the quantum of legal effort/drama in this one. We would need to know significantly more about the facts than is publicly known to adjust the quantums from similar cases into a semi-reliable estimate for this one.
Tyler: thanks for the post.
A couple questions I’d love to hear your thoughts on:
Would it be right to say that >90% of the legal risk to public figures and institutions was incurred prior to the FTX blow-up on November 8?
Do you expect that >30% of prominent EA figures will end up involved in the litigation process, regardless of what they post after November 8?
Do you expect that Effective Ventures (formerly: CEA) will end up involved in the litigation process, regardless of what they post after November 8?
Very off-the-cuff here, so forgive any sloppiness. 2 and 3 seem like empirical questions about which I don’t have enough information to even hazard a guess. I’d at best be stoking a similar kind of confusion my original post attempts to warn about.
Re 1, if we’re talking in the abstract about criminal or civil liability, then I think the answer is probably yes (with potential exceptions for situations where after-the-fact actions might give rise to liability for things like perjury and witness tampering, etc.).
But if we’re talking about cost in time, treasure and mental well-being associated with being a witness, or the knock-on effects for others and the system more generally, I think the answer is a strong no. Loose talk after a dispute has arisen can easily take a near-0% chance of someone being significantly involved (looking exclusively at pre-dispute facts) and make it a near certainty. And this would not necessarily be just for the individual or organization doing the talking. A simple but hopefully clear example would be where a central figure in the pre-dispute facts speaks post-dispute and identifies someone else as being significantly involved. The identified person would suddenly be on everyone’s radar even if they were on nobody’s radar before the statement, and that would be true even they weren’t actually involved (i.e., the post-dispute statement was false), and even if the false statement was an innocent mistake (a few examples of which I try to outline in the post—e.g., confusing post-dispute involvement with pre-dispute involvement, etc.).
Hope this is useful.
Very helpful, thanks Tyler!
If I understand you correctly, you’d agree that carefully saying true things, with extra attention to clarity (to avoid easy misreadings), would go a long way to reducing risk of (1).
I’m not sure I understand this comment completely, but I think I disagree with it. Speaking in the abstract because I don’t want to suggest I know anything about any specifics here, I think the best way to avoid legal risk (including but not limited to liability) based on pre-blow-up actions or statements in any situation like this is to stay silent after the fact and only say what is absolutely necessary through the formal legal process.
I am not Tyler but am a lawyer. Based on limited information, I think it likely that over 90% of legal exposure would be based on actions that happened on or before November 11 (let’s use the date of the bankruptcy filing). However, that is not equivalent to saying that what potential targets do from November 11 onward will affect the ultimate financial pain they end up bearing (damages, clawbacks, legal fees, intangible negatives) by no more than 10%. I think the latter statement is likely false, and is the question one has to consider when deciding on speech vs. silence.
The answer to #2 depends in part of your definitions of “prominent EA figures” or “involved in the litigation process.”
I would think less of Sullivan & Cromwell’s litigation team if EV and CEA are not involved in litigation in some capacity. I have zero inside knowledge, but way too much money went from FTX-affiliated entities and persons to EV and CEA not to make an attempt to clawback and to see what else might be there.
[Edited: “on or before November 11,” “November 11 onward.” Thanks to the person who messaged me to note those typos!]
Great questions! More succinct than what I said.
Why are so many people disagreeing with this?
My layman’s take agrees with all of Oliver’s points here.